The existing public opinion analysis models have hysteresis and inaccuracy, and the characteristics of individual choice affecting public opinion prediction are subjective and uncertain. In this paper, the ensemble empirical mode decomposition (EEMD) method and transformer attention mechanism are combined to propose a combined model EEMD⁃Transformer. The model decomposes the heat value of the original public opinion event by using EEMD decomposition technology. The decomposed data are extracted by feature extractor Transformer, and then predicted by a fully connected neural network. Taking the public opinion novel coronavirus pneumonia (corona virus disease 2019, COVID-19) as an example, the COVID-19 model is predicted by the trained model. The experimental results show that the research in this paper can accurately predict the trend of public opinion, which plays an important role in assisting the government and enterprises to guide the development of public opinion events.
YOUD D, CHENF J. Research on the prediction of network public opinion based on improved PSO and BP neural network [J]. Information Science,2016,35(8):156-161. DOI: 10.3969/j.issn.1002-1965.2016.08.027(Ch ).
[3]
潘新. 基于复杂网络的舆情传播模型研究[D].大连:大连理工大学,2010.
[4]
PANX. Opinion Spreading Models on Complex Networks[D]. Dalian:Dalian University of Technology,2010(Ch).
ZHOUY M, WANGB, ZHANGH C. Evolution analysis and modeling method of internet public opinions based on EMD[J]. Computer Engineering,2012,38(21):5-9.DOI: 10.3969/j.issn.1000-3428.2012.21.002(Ch ).
ZHAOJ H, WANK W. Research on the communication dynamics model of social network public opinion based on the SIS model [J]. Information Science,2017,35(12):34-38. DOI: CNKI:SUN:QBKX.0.2017-12-006(Ch ).
ZHANGY W, QIJ Y, WANB X, et al. Online public opinion risk warning based on bayesian network modeling[J]. Library and Information Service,2012,56(2):76-81 (Ch).
SUNJ C, ZHOUR, LIP Y, et al. Research on the prediction of network public opinion based on recurrent neural network [J]. Information Science,2018,36(8):120-124+129.DOI:CNKI:SUN:QBKX.0.2018-08-020(Ch ).
SHUY, ZHANGL L. Internet public opinions forecasting based on wavelet and artifical neural networks [J]. Information Science,2016,34(4):40-42+47. DOI: 10.13833/j.cnki.is.2016.04.008(Ch ).
[15]
魏德志,陈福集,郑小雪.基于混沌理论和改进径向基函数神经网络的网络舆情预测方法[J].物理学报,2015,64(11):52-59. DOI : 10.7498/aps.64.110503 .
[16]
WEID Z, CHENF J, ZHENGX X. Internet public opinion chaotic prediction based on chaos theory and the improved radial basis function in neural networks [J]. Acta Physica Sinica, 2015,64(11):52-59. DOI : 10.7498/aps.64.110503(Ch ).
DUZ T, XIEX Z. The establishment of public opinion forecasting and early-warning model with the methods of grey forecasting and pattern recognition [J]. Library and Information Service, 2013,57(15):27-33. DOI: 10.7536/j.issn.0252-3116.2013.15.004(Ch ).
ZENGZ M, HUANGC Y. Research on public opinion heat trend prediction model of emergent infectious diseases based on BP neural network [J]. Journal of Modern Information,2018,38(05):37-44+52 (Ch).
HUOF, GAOH Y, LIUC N, et al. Application of public opinion monitoring in control of major infectious diseases [J]. Occupation and Health,2013,29(23):3205-3206+3209.DOI: 10.13329/j.cnki.zyyjk.2013.23.014(Ch ).
[25]
HUANGN E, SHENZ, LONGS R,et al. The empirical mode decomposition and the hilbert spectrum for nonlinear and non-stationary time series analysis [J]. Proceedings of the Royal Society of London,1998, 454:903-905. DOI: 10.1098/rspa.1998.0193 .
[26]
WUZ H, HUANGN E. Ensemble empirical mode decomposition: A noise⁃assisted data analysis method [J]. Advances in Adaptive Data Analysis,2009,1(1):1-41. DOI: 10.1142/S1793536909000047 .
[27]
ZHENGK, LUOJ F, ZHANGY, et al. Incipient fault detection of rolling bearing using maximum autocorrelation impulse harmonic to noise deconvolution and parameter optimized fast EEMD [J]. ISA Transactions,2018,89:256-271. DOI: 10.1016/j.isatra.2018.12.020 .
[28]
CHENGY, WANGZ W, CHENB Y, et al. An improved complementary ensemble empirical mode decomposition with adaptive noise and its application to rolling element bearing fault diagnosis [J]. ISA Transactions,2019,91:218-234. DOI: 10.1016/j.isatra.2019.01.038 .
[29]
ASHISHV, NOAMS, NIKIP, et al. Attention is all you need [DB/OL].[2020-05-02].
[30]
HANT, LIUC, WUL J, et al. An adaptive spatiotemporal feature learning approach for fault diagnosis in complex systems [J]. Mechanical Systems and Signal Processing, 2019,117:170-187. DOI: 10.1016/j.ymssp.2018.07.048 .